This study evaluated the efficacy of a group cognitive treatment for pathological gambling. Gamblers, meeting DSM-IV criteria for pathological gambling, were randomly assigned to treatment (N=34) or wait-list control (N=24) conditions. Cognitive correction techniques were used first to target gamblers' erroneous perceptions about randomness, and then to address issues of relapse prevention. The dependent measures used were the DSM-IV criteria for pathological gambling, perceived self-efficacy, gamblers' perception of control, desire to gamble, and frequency of gambling. Post-treatment results indicated that 88% of the treated gamblers no longer met the DSM-IV criteria for pathological gambling compared to only 20% in the control group. Similar changes were observed on all outcome measures. Analysis of data from 6-, 12- and 24-month follow-ups revealed maintenance of therapeutic gains. Recommendations for group interventions are discussed, focusing on the cognitive correction of erroneous perceptions toward the notion of randomness.